Abstract:
Abstract: Plant's health status is one of the most important indicators to evaluate restoration projects and environmental evolution of degraded land, particularly in mine restoration areas. Most studies on vegetation cover changes did not consider the phenological characteristics and health status in remote sensing monitoring, although optical remote sensing technology has been available since the 1980s. Taking a basalt stone pit in Qitaihe City, Heilongjiang Province, China as the research area, this study aims to detect the health status of corn crops using remote sensing. Thirteen images were captured from the Sentinel-2 satellite from April to October 2020, and then reconstructed by the time-series of Normalized Difference Vegetation Index (NDVI). Different fittings were selected, including the Savitzky-Golay (S-G) filtering, double logistic (D-L) fitting, and asymmetric Gaussian (A-G) filtering. Dynamic thresholds and curvatures were used, where the phenological indicators of corn crops covered the periods of emergence, jointing, tasseling, and maturity, as well as the interval length between phenological periods. An analysis was made on the difference of corn phenological characteristics between the mine restored and normal cultivated areas. Some remote sensing spectral indices were extracted, including the Red-Edge Inflection Point (REIP), NDVI, green NDVI (GNDVI), Meris Terrestrial Chlorophyll Index (MTCI), pigment specific simple ratio (PSSRA), Inverted Red-Edge Chlorophyll Index (IRECI), Modified Chlorophyll Absorption Ratio Index (MCARI) and vegetation coverage (FVC). A systematic evaluation was made on the feasibility and sensitivity of indices to distinguish between healthy crops and sub-health crops in the restoration area. The results showed that the key phenological periods of corn was delayed in the restoration treatment area, where the emergence period was delayed by 5-12 days, while the jointing period was delayed by 9-12 days, and the maturity period was delayed by 21-22 days, compared with those in the normal cultivated areas. The phenological difference of crops gradually emerged as the crop grew. Remote sensing indicators, including the REIP, NDVI, GNDVI, PSSRA, IRECI, and FVC, can effectively distinguish the crops spectral features and the health status difference between the mine restoration and normal cultivation areas. The GNDVI received the lowest standard deviation for the restoration area (0.093 5 for area A, and 0.056 2 for area B), indicating the most stable indicator of the overall health status of crops. In PSSRA amplitude interval of pixels, the lower limit in the C area was 12.61% and 44.72% higher than the upper limit in the A and B areas, respectively, indicating the most sensitive among the indicators. The spectral features were related to the comprehensive influence of physiological difference between restoration and normal cultivation areas, including the leaf green content, leaf area index, and leaf water content of corn crops. This finding can provide an insightful understanding to rapidly evaluate the mine restoration using remote sensing technologies.